UV photoelectron spectra and DV-X.alpha ... - ACS Publications

Oct 1, 1987 - Albert Padwa, James P. Snyder, Erin A. Curtis, Scott M. Sheehan, Kimberly J. Worsencroft, and C. Oliver Kappe. Journal of the American ...
0 downloads 0 Views 940KB Size
Inorg. Chem. 1987, 26, 3406-3409

3406

by thermodynamic arguments: the greater flexibility of AT vs G C regions, the steric effects of the guanine amino group, the differences in potential in the G C and AT regions, etc. Alternatively, the apparent intercalative binding of TMpyP(4) and group I metalloporphyrins contrasted with the absence of such binding by group I11 porphyrins is explained by either a thermodynamic/steric argument (the presence of axial ligands) or a kinetic/steric argument (inability to achieve a planar structure for TMpyP(2), etc). As stated above, our major objectives here involved evaluating the effects of metalloporphyrins on D N A properties to identify areas where fruitful studies might be conducted with oligonucleotides. However, the apparent differences in effects of group I porphyrins (TMpyP(4), NiTMpyP(4), and PdTMpyP(4)) when combined with the selectivity of TMpyP(4) for 5'-CG-3' over 5'-GC-3' sites have led us to wonder if all effects are not best explained by primarily a thermodynamic/steric argument. The highly charged porphyrins appear to be binding by many competitive modes. If such subtle differences between the group I porphyrins and between sequences lead to differences in binding, then perhaps major changes (TMpyP(4) to TMpyP(2), Nmethylpyridiniumyl to trimethylaniliniumyl) could make intercalative binding unfavorable. (Again, we repeat that we do not claim that intercalative binding has been established beyond question). For example, steric effects of the N-methyl groups of TMpyP(2) could easily prevent intercalation from being thermodynamically possible. Likewise, intercalated TMAP, with its positive groups ca. 1.5 8, further out from the porphyrin center than TMpyP(4), might not be able to form as favorable interactions with the phosphate groups of DNA as does intercalated TMPYP(4).

We believe that this thermodynamic/steric model explains the interaction of complex molecules, like the porphyrins, with DNA and offers an alternative to the "temporarily planar" kinetic explanation for additional reasons. First, complex molecules such as n o g a l a m y ~ i n ~and z ~ naphthalenediimide~?~*~~ ~ which have bulky groups on opposite sides of the planar ring system and which cannot become planar, bind to D N A by intercalation. Second, model building studies indicated that porphyrins, even with the pyridyl groups in the plane of the porphyrin ring, still require significant disruption of the double helix to allow intercalation. It is clear that a complete understanding of the factors influencing binding of these intriguing porphyrin species must await further information on binding modes and additional kinetic measurements. In particular, studies with oligonucleotides may allow us to differentiate between the models. Oligonucleotides undergo large amplitude dynamic motions, including complete strand separation. These motions would easily allow TMpyP(2) to intercalate if only a kinetic barrier prevented insertion of the porphyrin. Acknowledgment. This work was supported by NIH Grant GM 29222 to L.G.M. and NSF Grant DBM 8603566 to W.D.W. The Col E, CCS DNA was a gift from Dr. Robert L. Jones. The 360-MHz N M R spectrometer was purchased with partial support from the NSF through a departmental grant to Emory University. (52) Arora, S. K. J . Am. Chem. SOC.1983, 205, 1328. (53) Collier, D. A.; Neidles, S.; Brown, J. R. Biochem. Pharmacol. 1984, 33, 2877. (54) Yen, S. F.; Gabbay, E. J.; Wilson, W. D. Biochemistry 1982, 21, 2070. (55) Hopkins, H. P.; Stevenson, K. A.; Wilson, W. D. J . Solution Chem. 1986, 15, 563.

Notes Contribution from the Dipartimento di Chimica Inorganica, Metallorganica ed Analitica, Universitl di Padova, Padova, Italy, Istituto di Chimica e Tecnologia dei Radioelementi del CNR, Padova, Italy, and DiPartimento di Chimica Inorganica e Struttura Molecolare, Universita di Messin% Messin% Italy

UV Photoelectron Spectra and DV-Xa Calculations on RhZ4+ Formamidinate Complexes Gian Andrea Rizzi,la Maurizio Casarin,Ib Eugenio Tondello,Ia Pasquale Piraino,Ic and Gaetano Granozzi*Ia Received March 2, 1987 Complexes containing the Rh24+core have been extensively investigated from both the synthetic2 and spectroscopic2 points of view, and their possible applications have been o ~ t l i n e d . Most ~ of these studies have been directed toward complexes containing four carboxylate groups as bridging ligands. Several theoretical studies (using both MS-Xu and minimal basis-set ab initio methods) on the nature of the Rh-Rh bond in carboxylates have been reported! They are all in agreement with the existence of a single Rh-Rh bond, but different electronic

configurations for the tetraformate Rh2[O2CHI4complex (1) have been obtained according to the different theoretical approaches ab initio,4b & 3 2 ~ * 4 6 * 2 ~ 2 ) . However, (MS-Xa,4a &4&*4,3*2; no experimental determination of one-electron levels by gas-phase photoelectron (PE) spectroscopy has been reported so far because of the thermal decomoosition of the carboxvlate dimers in the ionization chamber. T'o our knowledge, the &que PE literature data regarding Rh24+complexes refer to the complex Rh2(mhp)4 ~ T (mhp = 2-oxo-6-methylpyridinate), where a C T ~ T ~ ~configuration was proposed in order to nicely fit the experimental data.5 Recently, complexes containing formamidinate ligands have been prepared and structurally characterized6 and the attempt to record their PE spectra has been successful. In this contribution we report the UV excited P E spectra of Rh2[HC(NR),12[02CCFJ2 (2) and Rh2[HC(NR)2]4(3) (R = p-tolyl) coupled with the results of first-principle discrete variational (DV) X a quantum-mechanical calculations7 carried out on the Rh2[HC(NH),],[O,CH], (2a) and Rh2[HC(NH)2]4(3a) models.

Experimental Section Synthesis. The samples of Rh2[HC(NR)~]2[02CCF~lz(H20), (2c) and €UI~[HC(NR)~]~ (3) (R = p-tolyl) were synthesized according to the published procedures.6 After crystallization, their purity was checked by IR and N M R measurements.

(1) (a) University of Padova. (b) CNR of Padova. (c) University of

Messina. (2) (a) Boyar, E. B.; Robinson, S. D. Coord. Chem. Reo. 1983,50, 109. (b) Felthouse, T. R. Prog. Inorg. Chem. 1982, 29, 73. (3) (a) Hughes, R. G.; Bear, J. L.; Kimball, A. P.Proc. Am. Assoc. Cancer Res. 1972, 13, 120. (b) Howard, R. A,; Kimball, A. P.; Bear, J. L. Cancer Res. 1979, 39, 2568. (4) (a) Norman, J. G., Jr.; Kolari, H. J. J . Am. Chem. Soc. 1979, 101, 5256. (b) Nakatsuji, H.; Ushio, T.; Kanda, K.; Onishi, Y . ;Kawamura, T.; Yonezawa, T. Chem. Phys. Lett. 1981, 79, 299.

0020-166918711326-3406$01.50/0

(5) Berry, M.; Garner, C. D.; Hillier, I. H.; McDowell, A. A,; Clegg, W. J . Chem. SOC.,Chem. Commun. 1980, 494. (6) (a) Piraino, P.; Bruno, G.; Tresoldi, G.; Lo Schiavo, S.; Zanello, P. Inorg. Chem. 1987, 26, 91. (b) Piraino, P.; Bruno, G.; Tresoldi, G.; Lo Schiavo, S.; Zanello, P. Inorg. Chem., in press. (7) (a) Averill, F. W.; Ellis, D. E. J . Chem. Phys. 1973, 59, 6411. (b) Rosen, A.; Ellis, D. E.; Adachi, H.; Averill, F. W. J . Chem. Phys. 1976, 65, 3629 and references therein. ( c ) Trogler, W. C.; Ellis, D. E.; Berkowitz, J. J . Am. Chem. Sot. 1979, 101, 5896.

0 1987 American Chemical Society

* ~ ~ * ~

Inorganic Chemistry, Vol. 26, No. 20, 1987 3407

Notes Table I. DV-Xa Results on Rh2[HC(NH)2]4 population, % 2 Rh MO 4a2, (LUMO)

-e,

eV

1.58 2.04 3.08 3.99 4.15 4.63 4.76 5.30 5.88 6.79 7.05 7.94 7.95 8.20 8.56

S

P

d

8 N

15

5

67

12 38 1 100

62 95

3

89 72 2 31

2 4

17 37

12 H 1

1

3

25

75 18

4 c

100 5 7 67 69 92 49 46 33 52

3 1 10

3 2 18

1 12 54 14 48

5 18 16

dominant charactera u* Rh-Rh 6* Rh-Rh, a Rh-N ab a* Rh-Rh N nb 6 Rh-Rh N nb T Rh-Rh u Rh-Rh u ligand a Rh-N b N nb (T Rh-N b a ligand Rh-N b a ligand

OLegend: b = bonding; ab = antibonding; nb = nonbonding.

Spectra. He I and He I1 excited PE spectra were measured on a Perkin-Elmer PS- 18 spectrometer modified for He I1 measurements by inclusion of a hollow cathode discharge lamp giving a high output of He I1 photons (Helectros Developments). The ionization energy (IE) scale was calibrated by reference to peaks due to admitted inert gases (Xe-Ar) and to the He Is-' self-ionization. A heated inlet probe system was adopted at 290 OC (3) and 210 "C (2c). Under the experimental conditions in the ionization chamber complex 2c loses the axial water molecules, giving rise to 2, as clearly monitored by the presence (in the 50-70 OC temperature range) of the PE peaks due to the free H 2 0 molecule. Theoretical Method. Hartree-Fock-Slater (HFS) discrete variational (DV-Xa) calculations7 were performed on a VAX 8600 computer. Numerical atomic orbitals obtained for the neutral atoms (through 5p for Rh and 2p for C, N, and 0) were used as basis functions. Due to the size of the investigated systems, orbitals 1s-4p on Rh and 1s on C, N, and 0 were treated as a frozen core in the molecular calculations. A self-consistent multicenter-multipolar (SCM) representation of the density is introduced according to a procedure described elsewhere.* Five radial degrees of freedom were allowed in the expansion of the density in addition to radial atomic densities. The Gaspar-Kohn-Sham value9 for the exchange scaling parameter has been adopted ( a = 2/3). Atomic orbital populations were computed by using the Mulliken analysis.10 To evaluate the magnitude of electronic relaxation associated with the removal of one electron from the various ground-state MOs, IE calculations were carried out by using Slater's transition-state formalism (TSIE)." Calculations have been carried out on 2a and 3a models by using the geometrical parameters obtained from the solid-state structural determinations of 2~~~and 36band assuming perfect Ch (for 2a) and D4* (for 3a) structures (see Figure 1 for the axis convention).

Results and Discussion In Table I the charge density analysis of the outermost MOs of 3a is reported. The dominant character has been assigned on the basis of symmetry considerations and by an accurate analysis of the relative wave functions. The electronic configuration describing the metal-metal interaction is &r462?r*46*2. A single Rh-Rh bond is then predicted because only the u component (5alg) has an empty antibonding counterpart (the 4a2, LUMO). The population analysis shows that 5s and 5p Rh orbitals are slightly involved in the metal-metal bonding, the only significant contribution arising from the 5s orbital in the u and u* MOs. It is interesting to notice that some MOs, involved in the metal-metal interaction, have also a relevant metal-ligand, either bonding or antibonding, character (Table I). For example, both lbluand 2bl, MOs have a 6* M-M character,12 but the outermost (8) Delley, B.; Ellis, D. E. J. Chem. Phys. 1982, 76, 1949. (9) (a) Gaspar, R. Acta Phys. Acad. Sci. Hung. 1954, 3, 263. (b) Kohn, W.; Sham, L. J. Phys. Rev. 1965, 140, A1133. (10) Mulliken, R. S. J. Chem. Phys. 1955, 23, 1833. (1 1) Slater, J. C. Quantum Theory of Molecules and Solids. The SevConsistent Field for Molecules and Solids; McGraw-Hill: New York, 1974; Vol. 4.

z

--x

2a

Figure 1. Schematic views of investigated molecules. The axis system is also reported.

one (2bl, HOMO) also has an antibonding Rh-N ?r character, while the innermost 1blu has a T Rh-N bonding character. Other MOs contributing to the Rh-ligand bonding are 4a,, and 4bl, ( u in nature). Orbitals of the eg and e, type do not contribute at all to the metal-ligand interaction and are localized either in the metal or in the ligand framework. A correlation diagram between the ground-state DV-Xa levels of the formamidinate complexes 2a and 3a and those obtained from a similar calculation on the Rh2[02CH], tetraformate complex (1) is reported in Figure 2. Simple electronegativity arguments about the bridging ligands (oxygen vs. nitrogen) can explain the progressive destabilization of all the valence levels on going from 1 toward 3a. In all three cases the presence of a single metal-metal bond is predicted, in agreement with previous theoretical s t ~ d i e s .It~ is interesting to note the enhancement of the energy separation between the ?r* and 6* components on going from 1 toward 3a. This effect can be easily explained by taking into account the (12) The outermost 2blu MO has the largest localization on the metallic centers, and it has been taken as representative of the 6* interaction.

3408 Inorganic Chemistry, Vol. 26, No. 20, I987

Notes

Table 11. IE and TSIE Values (eV) of Rh2[HC(NRZ)l4and Rh2[HC(NRz)]z[02CCF3]2(R = p-Tolyl) band A

B C

D E E'

F G G'

Rh,lHC(NR,)l, . -. - . assignt 2b,, Sei-+ la,, 2b2, 4e, 6e, 5a,,

IE 5.69 6.63 7.36 7.84 8.56 9.43 11.22 13.50 14.74

-

--

Rh, IHC(NR,)l ,IO,CCF,17 . -,._. _._

TSIE

band A

4.85 6.39, 6.43 7.06 7.05 7.8 1 8.55

B C

D E

E' F F' G

IE

TSIE

assignt

6.68 7.21 7.70 8.37 9.12 9.85 10.63 11.57 14.00

lobi Sa2 9az + 9bl (13al 12al + l l b 2

+

6.48 7.98 7.41, 7.51 8.48, 8.51, 8.50, 8.76

+ 8bJ

lla,

9.58

eV

.I.

/

4a2u LUMO

'*

-2

-3

.4

Ligand

s Ligand x

-5 0

-6

-7

5

6

7

8

9

10

11

12

13

14

15

Figure 3. He I/He I1 excited PE spectra of Rhz[HC(NR)2]4 (R = p-to1yl).

-8

An interesting point that emerges from the calculations is that are not the Se, (a* M-M) and 6% ( a M-M) orbitals of 3a (D4,,) a2,bl and e, al,b2 significantly split in 2a (by symmetry e, in C,) so that the reduction of symmetry is only a formal matter as far as the M-M interaction is concerned. This theoretical prediction is fully confirmed by the PE spectra (see below). The He I and He I1 excited PE spectra of 3 and 2 are reported in Figures 3 and 4,where bands have been alphabetically labeled. The corresponding IE values and the assignments of the lowest IE bands (the most interesting ones, since they are related to ionizations from orbitals involved in the Rh-Rh interactions) are reported in Table 11. In the same table the DV-Xa-computed TSIB are also reported. The presence of the ptolyl substituents greatly complicates the PE spectra since the outermost r ionizations from the aromatic rings give rise to bands starting from ca. 8.5 eV. The assignments reported in Table I1 refer t o ionizations from the model compounds 2a and 3a, with the contributions from the p-tolyl substituents omitted.I3J4 According to the calculations (Tables I and 11), band A of the H e I PE spectrum of 3 (Figure 3) is assigned to ionization from the 6* 2b1, HOMO. Its relative intensity with respect to the following band B is remarkably enhanced in the He I1 spectrum. If we assume the usual empirical arguments about the relative

-

u

Figure 2. Correlation diagram between one-electron energy levels of l,, 2a, and 3n.

above-mentioned Rh-bridging ligand a-antibonding nature of the 6* H O M O (see Table I): the better energy matching between the 6* metallic orbitals and the ligand-based a orbitals, when the formamidinate ligand is considered, gives rise to a stronger Rhligand antibonding interaction, so destabilizing the resulting 6* MO. Another interesting trend along the series 1 2a 3s is the progressive destabilization of the 6 component with respect to the r one, leading to a crossing between the two levels on passing from 1 to 2a (Figure 2). The steric hindrance due to the p-tolyl substituents Can explain the lack of axial ligands in the crystal structure of 3,6bat variance with the case for 1 and 2c, whose crystal structures show axial water molecules.2a,6a However, the strong destabilization of all levels (and in particular of the 4az, LUMO) in the 3a complex suggests that also an electronic factor acts adversely on the axial coordination. In fact, in a previous MS-Xa study of 1, Norman and K 0 1 a r i ~pointed ~ out that the Rh-H20(ax) interaction is basically of u type and arises mainly from the combination of oxygen lone pairs with the LUMO 4a2, (u* Rh-Rh). Therefore, an unfavorable energy matching between the LUMO and the orbitals of the axial donor can concur to explain the scarce affinity of complex 3 toward axial ligands.

- -

-

(13) For example, band E of the spectrum of 3 has been assigned to the ionization from the 6e, M O (Table 11) but it also contains ionizations from the aromatic rings. (14) The CF, substituents only produce shifts on the outermost PE bands of the spectrum of 2 since the CF,-localized ionizations occurr at higher

IEs.

3409

Inorg. Chem. 1987, 26. 3409-3413

that in 2 we have only half of the aromatic rings (whose ionizations contribute to band E) that were present in the tetraformamidinate complex 3. Very important experimental evidence is related to the change of the relative intensity between bands B and C on passing to the H e I1 spectrum (Figure 4): the large falloff of intensity of band B is diagnostic of localization on the ligand of the corresponding MO." As in the previous case, band A is safely assigned to ionization from the lobl (6*) HOMO on the basis of the calculations. The assignment of bands B and C, however, needs some discussion. We have already pointed out that the symmetry reduction on passing from 3 to 2 does not produce significant energy separation on the a type metal-metal orbitals (see Figure 2 and TSIEs in Table 11). According to the TSIE values obtained for the model 2a, we should assign band B to the a* (Rh-Rh) (9a2,9bl) MOs (over 90% localization on the metals) and band C to the ligand-based 8a2 M O (over 95% on the formamidinate ligand). This assignment, however, is in contrast with two experiment1 observations: (i) band B is less intense than band C in the He I spectrum; (ii) the already mentioned decrease of the relative intensity of B vs. that of C on passing to the He I1 spectrum. Both these points give support to the opposite assignment, Le. band B to 8a2 and band C to 9a2 and 9bl MOs. This failure of the theoretical prediction could be ascribed to the neglect of the CF, groups in the model adopted for the cal~ulations.'~ We take band D to represent the ionizations of four MOs, namely 13al (6 bonding Rh-Rh), 12al and 12b2 (a bonding Rh-Rh), and 8b1 (localized to an equal extent on both formamidinate and carboxylate ligands). Finally, in analogy with the assignment of 3, we propose the assignment of the 1l a l ( u bonding Rh-Rh) M O to the weak shoulder E'. In conclusion, we want to stress here that the present analysis of the UV-PE data on Rh;' formamidinate complexes furnishes 6**) with that an electronic configuration ( ~ * 7 r ~ t j ~ a * ~identical obtained by a previous UV-PE study on the quite different Rh2(mhp), ~ o m p l e x .On ~ the other hand, all the available X a type theoretical results4* (the most rigorous treatment for such complicated molecules to date) are in agreement in predicting similar configurations in both carboxylate and formamidinate complexes. We could suggest, then, that this bonding scheme can be considered as peculiar to the Rh?' arrangement. Acknowledgment. Financial support for this study from the Minister0 della Pubblica Istruzione (Rome) is gratefully acknowledged. Registry No. 2, 109976-30-9;3, 108149-48-0.

i p I If

eV 5 6 i 8 9 Io ii 1'2 i ' 3 - . . Z X Figure 4. He I/He I1 excited PE spectra of Rh2[HC(NR)2]2[02CCF~]2

(R = p-tolyl). intensity variations on passing from the He I to the He I1 ionizing ~ o u r c e ,the ' ~ observed behavior can be interpreted either as a major involvement of the d metal AOs with the orbital giving rise to band A or as an indication of some component of band B arising from a ligand-localized MO. The latter hypothesis is well in tune with the theoretical results, which predict, at ca. 1.5 eV higher IE than the ionization from the HOMO, two quasi-degenerate ionizations from the 5e, (a*Rh-Rh, 95% localized on metals) and lalu (100% nitrogen-based a ligand) MOs (see Table I). The presence of two components under band B is also in agreement with the splitting observed in the corresponding band of the PE spectrum of the less symmetrical complex 2 (Figure 4). The low-intensity band C (which has been confirmed by running several expanded-scale He I spectra) can be tentatively assigned to the ionization of the 6 (Rh-Rh) 2b2, MO. Its low intensity agrees well with that observed for the same ionization in the spectrum of Rh2(mhp)4.5 The assignments of the inner bands are complicated by the superposition of the ptolyl group ionizations. We can just propose the following plausible assignments on the basis of the results of the theoretical calculations. The computed energy separations between the 4e, (ligand u/a nonbonding MO), 6e, (a Rh-Rh), and 5a1, ( u Rh-Rh) MOs lead us to ascribe their ionizations respectively to shoulder D, band E, and shoulder E'. The H e I P E spectrum of the mixed complex 2 (Figure 4) is noticeably different from that just discussed: all the bands are shifted by about 1 eV toward higher IEs, as already anticipated by our theoretical results (see Figure 2 and TSIEs in Table 11). Moreover, band B of the spectrum of 3 is now split into two well-resolved components (B C) and band D, a shoulder of band E in the previous case, is now a resolved band with an intensity comparable to that of band E. In this regard we must remember

(17) Actually, when one deals with a less symmetrical molecule, such as 2, the errors induced by adopting a simplified model may not be balanced. In this case, the 9a2 and 9bl MOs have a larger localization on the carboxylate ligands than the 8a2 MO, so that the neglect of the large electron-withdrawing effect of the CF, substituents produces a shift toward lower IEs of the 9a, and 9b, MOs.

Contribution from the Department of Chemistry and Biochemistry, University of California, Los Angeles, California 90024-1569

Synthesis and Molecular Structures of closo -3-(q6-CH3C6HS) -3,1,2-FeC2B9H1 and closo - 3 4 q6-1,4-(CH3)2C6H4)-3,1,2-FeC2B9H,I

+

Han Chyul Kang,,Carolyn B. Knobler, and M. Frederick Hawthorne*

(15) In fact, on the basis of the Gelius model for the molecular photoioni-

Received April 16, 1987

zation cross sections,16we expect a marked decrease in relative intensity for those bands containing ligand-localized ionizations with respect to metal nd-based ones on passing from the He I to the He I1 excitation source. (16) Gelius, U.In Electron Spectroscopy; Shirley, D. A,, Ed.; North-Holland: Amsterdam, 1972; p 311.

0020-1669/87 , ,/ 1326-3409$01.50/0 I

The chemistry of mixed-sandwich transition-metal complexes of the type [(&arene)M(~~~-Cp)] has been well documented in literature.' In view of the close structural and electronic similarity Q 1987 American Chemical Societv -